Dollars and sense
The bar at the Eden Roc hotel in Miami Beach, Fla., seemed an unlikely place for a conversation on wireless data deployments, and paper napkins and peanuts were not exactly the best tools to illustrate the point my friend, Peter Thornycroft (of Aruba Networks) was making. Nonetheless, Peter managed to enthusiastically get his point across about how improvements to wireless data networks would change the way people viewed wireless.
In his hands, those peanuts and napkins transformed into a portrayal of the breadth of possibilities the then nascent 802.11n standard offered. He showed my colleague and me how multipath signals could be used to double the connectivity to a wireless client. Combined with some optimizations to existing algorithms, and doubling the amount of in-air bandwidth from 20 MHz to 40 MHz, we could visualize how clients could connect at or faster than 300 Mbps.
That conversation happened in November 2004, nearly five years before the IEEE ratified the 802.11n standard. Five months later, in May 2005, we installed the first campus-wide wireless network at Brandeis University in Waltham, Mass., just outside Boston. We had been looking at a wireless upgrade for some time, and with 802.11n years away from reality, I made the decision — after reviewing multiple vendors’ wireless solutions — to install Aruba’s 802.11 a/b/g access points across 100 buildings on the campus. The network we installed gave us sufficient coverage to call it an auxiliary network; everyone could connect and obtain signal strength of at least two bars. However, while connectivity was good, it was not quite sufficient to justify dumping our wired connections.
What we didn’t foresee was the consumer wireless revolution that was just emerging at that time, but which soon would spread like wildfire. Students overwhelmingly wanted wireless connections instead of wired ones. In fact, they chose to use the wireless network, despite its obvious shortcomings, over the dedicated wired port available to them in each dorm room and study cubicle. Though we had set up the wireless network purely as an add-on to the wired network, suddenly most students were using it as their primary form of access. When our help desk started to receive daily complaints that two bars were inadequate and about slow network access, we realized it was time for action. Thus began the process of filling coverage holes and looking towards a wireless future.
My personal wireless revolution began in the summer of 2008 when I experienced first-hand the speed of 802.11n. It was a transformative experience, and I knew in an instant that it would change how I deployed networks.
With the 802.11n standard now ratified, enterprises at last have a choice of edge-networking technology — 802.11n or Ethernet. With the majority of deployed Ethernet edge ports stuck at 100 Mb and offering zero security, wireless, in contrast, has a lot of advantages going for it.
Wireless networks need to be assessed differently than traditional wired networks. Wireless access points are only 50% of the equation. If most of your clients are desktop systems that are well served by an existing (and fairly modern) wired network, then you may not have many clients connecting. On the other hand, if your existing wired clients are older or not well-served — or if the switch gear is due for a refresh — then wireless holds many opportunities for you.
On the other hand, if your users are mostly mobile and connecting using recently purchased equipment — including smart phones and PDAs — expect that they already have switched to wireless as their primary network and will expect the infrastructure to meet their expectations.
A site survey is an essential prerequisite to a network deployment. But just mentioning a site survey can cause IT directors to go into economic shock. Professional site surveys can cost tens of thousands of dollars and they only provide a snapshot of the point in time at which they were done. Learning to do your own site survey will give you the ability to plan your network and adapt to future changes to your office space. I could dedicate a whole article just to site surveys, but here are a few quick tips you can use right away:
- Program an AP with a testing SSID and put it on a 50-foot cable so you can move it as needed.
- Think of your AP as a light bulb and position it to eliminate the shadows — even a few feet can make a big difference in coverage.
- Use a real mobile device to take measurements because a computer simulation can’t replicate the actual end-user experience.
We recently built a $25 million science center and found during construction that we were over budget. I successfully made the case that by adding 802.11n access points to the project we could scale back the wired network, compensate for some lost drops and prepare for future expansion while still saving money. And that’s exactly what happened — we saved hundreds of thousands of dollars in construction changes while enhancing user mobility in the process.
In another renovation project we faced the possibility of remediating improperly installed Cat-5 wiring to the tune of $300,000. We chose to take a proactive measure and remove the wiring in order to install a high-density 802.11n wireless network. The cost of the new wireless network was one-third the cost of fixing the wired network.
Greenfield projects offer the largest possible cost-savings opportunities. A conservative approach — which maintained wired infrastructure where it was considered mandatory — demonstrated a 50% cost savings over traditional wired deployments.
To further exploit the cost advantage of wireless, I created a new standard for the campus that elevates the wireless network to a level where it could be the primary access layer. That standard includes the following requirements:
- Deploy 802.11 a/b/g/n access points as the new standard for all new construction and renovated spaces.
- Achieve a -62 dBi signal on 5 GHz for primary occupied spaces.
- Reduce wired port deployments by 50%.
Using this strategy we were able to deploy the highest quality network in new buildings and add a tremendous performance improvement in renovated spaces.
No network is going to be all-wireless because you still need ports to connect certain devices such as access points and servers. However, users are moving away from wires in favor of wireless. Even building-control networks for heating, ventilation, air conditioning and lighting are moving toward wireless — and the infrastructure needs to follow.
The basic economic advantages of wireless will drive more devices toward the technology in the future. Those who take advantage of the usability and cost-saving opportunities wireless has to offer over wired networks, by building a strong wireless infrastructure, will position themselves well for the future. Digging in deep and examining where the real costs of networking lay will reveal that wireless will save you large sums in the long run and make users happy in the process.
John W. Turner is the director of networks and systems at Brandeis University. This article was contributed by Aruba Networks.